||Introduction: Alzheimer’s disease (AD) is a devastating, progressive, irreversible
brain disorder. Previous research has identified genes associated with the risk of
developing AD. Variations in the Apolipoprotein E (APOE) gene show the largest effect
size, with the ε4 isoform associated with highest risk. Genome Wide Association Studies
(GWAS) have found other genes associated with AD, yet none with effects as large as
APOE. Because AD diagnosis is often preceded by a long period of cognitive decline,
we investigated the relationship between previously determined AD risk genes and cognitive
decline to determine whether we could detect individuals at risk of imminent decline
and at a high priority for clinical intervention.
Methods: Cognitively healthy participants from the Duke “MURDOCK” study based in Kannapolis,
NC participated in the study. They were aged 55+, had contributed DNA, and undergone
two waves of cognitive assessments 4 years apart (n=713). An AD genetic risk score
(AD-GRS) was derived for each individual based on the known 9 AD genes from recent
meta-analyses. APOE was modeled separately. Scores were based on number of risk alleles
and the associated odds ratio for each gene. To determine optimal measure of cognitive
decline, the available cognitive tests were evaluated individually and in three different
composite measures (1.Global; 2.Learning/Memory; 3.Memory-Weighted). Post-hoc analyses
evaluated interactions between AD-GRS, APOE risk-score, combined risk-score, cognition,
and cognitive decline as measured by composites and individual assessments.
Results: APOE risk-score was associated with cognitive decline as measured by all
of the composite measures. APOE risk-score was most highly associated to the newly
constructed Predict Composite (composed of assessments most associated to risk genes),
followed by the Memory-Weighted, the Learning/Memory, and lastly the Global Composite.
APOE risk-score was associated with individual assessments except delayed recall.
The AD-GRS was not associated with cognitive decline but associated with baseline
cognition as measured by composites weighting memory. The combined risk-score was
less associated with cognitive decline than APOE alone.
Conclusions: APOE was associated with cognitive decline as best captured by the composites
that weighted memory. Although associated with AD, the other risk genes were not associated
with cognitive decline, yet are related to baseline cognition best captured by composites
weighting memory. This suggests that for identifying individuals at risk of cognitive
decline, focusing on APOE will be more useful than other AD risk alleles, and that
the optimal composite for capturing change associated with AD appears to be one that
is weighted with memory.